drawing disorders in alzheimer's disease

7/26/2019 Drawing Disorders in Alzheimer's Disease

1/23

See discussions, stats, and author profiles for this publication at:https://www.researchgate.net/publication/301592114

Drawing Disorders in Alzheimer's Diseaseand Other Forms of Dementia

Article in Journal of Alzheimer's disease: JAD April 2016

Impact Factor: 4.15 DOI: 10.3233/JAD-160009

READS

19

2 authors:

Luigi Trojano

Second University of Naples

226PUBLICATIONS 2,867CITATIONS

SEE PROFILE

Guido Gainotti

Catholic University of the Sacred Heart

337PUBLICATIONS 9,884CITATIONS

SEE PROFILE

Available from: Luigi Trojano

Retrieved on: 26 April 2016
https://www.researchgate.net/publication/301592114_Drawing_Disorders_in_Alzheimer%27s_Disease_and_Other_Forms_of_Dementia?enrichId=rgreq-6f6baea8-4e64-4b77-a198-380e0368005a&enrichSource=Y292ZXJQYWdlOzMwMTU5MjExNDtBUzozNTUwODUzMDY2MTM3NjFAMTQ2MTY3MDM0MjU5OQ%3D%3D&el=1_x_2https://www.researchgate.net/publication/301592114_Drawing_Disorders_in_Alzheimer%27s_Disease_and_Other_Forms_of_Dementia?enrichId=rgreq-6f6baea8-4e64-4b77-a198-380e0368005a&enrichSource=Y292ZXJQYWdlOzMwMTU5MjExNDtBUzozNTUwODUzMDY2MTM3NjFAMTQ2MTY3MDM0MjU5OQ%3D%3D&el=1_x_2https://www.researchgate.net/publication/301592114_Drawing_Disorders_in_Alzheimer%27s_Disease_and_Other_Forms_of_Dementia?enrichId=rgreq-6f6baea8-4e64-4b77-a198-380e0368005a&enrichSource=Y292ZXJQYWdlOzMwMTU5MjExNDtBUzozNTUwODUzMDY2MTM3NjFAMTQ2MTY3MDM0MjU5OQ%3D%3D&el=1_x_3https://www.researchgate.net/publication/301592114_Drawing_Disorders_in_Alzheimer%27s_Disease_and_Other_Forms_of_Dementia?enrichId=rgreq-6f6baea8-4e64-4b77-a198-380e0368005a&enrichSource=Y292ZXJQYWdlOzMwMTU5MjExNDtBUzozNTUwODUzMDY2MTM3NjFAMTQ2MTY3MDM0MjU5OQ%3D%3D&el=1_x_3https://www.researchgate.net/profile/Luigi_Trojano?enrichId=rgreq-6f6baea8-4e64-4b77-a198-380e0368005a&enrichSource=Y292ZXJQYWdlOzMwMTU5MjExNDtBUzozNTUwODUzMDY2MTM3NjFAMTQ2MTY3MDM0MjU5OQ%3D%3D&el=1_x_5https://www.researchgate.net/institution/Second_University_of_Naples?enrichId=rgreq-6f6baea8-4e64-4b77-a198-380e0368005a&enrichSource=Y292ZXJQYWdlOzMwMTU5MjExNDtBUzozNTUwODUzMDY2MTM3NjFAMTQ2MTY3MDM0MjU5OQ%3D%3D&el=1_x_6https://www.researchgate.net/profile/Luigi_Trojano?enrichId=rgreq-6f6baea8-4e64-4b77-a198-380e0368005a&enrichSource=Y292ZXJQYWdlOzMwMTU5MjExNDtBUzozNTUwODUzMDY2MTM3NjFAMTQ2MTY3MDM0MjU5OQ%3D%3D&el=1_x_7https://www.researchgate.net/profile/Guido_Gainotti?enrichId=rgreq-6f6baea8-4e64-4b77-a198-380e0368005a&enrichSource=Y292ZXJQYWdlOzMwMTU5MjExNDtBUzozNTUwODUzMDY2MTM3NjFAMTQ2MTY3MDM0MjU5OQ%3D%3D&el=1_x_5https://www.researchgate.net/institution/Catholic_University_of_the_Sacred_Heart?enrichId=rgreq-6f6baea8-4e64-4b77-a198-380e0368005a&enrichSource=Y292ZXJQYWdlOzMwMTU5MjExNDtBUzozNTUwODUzMDY2MTM3NjFAMTQ2MTY3MDM0MjU5OQ%3D%3D&el=1_x_6https://www.researchgate.net/profile/Guido_Gainotti?enrichId=rgreq-6f6baea8-4e64-4b77-a198-380e0368005a&enrichSource=Y292ZXJQYWdlOzMwMTU5MjExNDtBUzozNTUwODUzMDY2MTM3NjFAMTQ2MTY3MDM0MjU5OQ%3D%3D&el=1_x_7https://www.researchgate.net/profile/Guido_Gainotti?enrichId=rgreq-6f6baea8-4e64-4b77-a198-380e0368005a&enrichSource=Y292ZXJQYWdlOzMwMTU5MjExNDtBUzozNTUwODUzMDY2MTM3NjFAMTQ2MTY3MDM0MjU5OQ%3D%3D&el=1_x_7https://www.researchgate.net/institution/Catholic_University_of_the_Sacred_Heart?enrichId=rgreq-6f6baea8-4e64-4b77-a198-380e0368005a&enrichSource=Y292ZXJQYWdlOzMwMTU5MjExNDtBUzozNTUwODUzMDY2MTM3NjFAMTQ2MTY3MDM0MjU5OQ%3D%3D&el=1_x_6https://www.researchgate.net/profile/Guido_Gainotti?enrichId=rgreq-6f6baea8-4e64-4b77-a198-380e0368005a&enrichSource=Y292ZXJQYWdlOzMwMTU5MjExNDtBUzozNTUwODUzMDY2MTM3NjFAMTQ2MTY3MDM0MjU5OQ%3D%3D&el=1_x_5https://www.researchgate.net/profile/Guido_Gainotti?enrichId=rgreq-6f6baea8-4e64-4b77-a198-380e0368005a&enrichSource=Y292ZXJQYWdlOzMwMTU5MjExNDtBUzozNTUwODUzMDY2MTM3NjFAMTQ2MTY3MDM0MjU5OQ%3D%3D&el=1_x_4https://www.researchgate.net/profile/Luigi_Trojano?enrichId=rgreq-6f6baea8-4e64-4b77-a198-380e0368005a&enrichSource=Y292ZXJQYWdlOzMwMTU5MjExNDtBUzozNTUwODUzMDY2MTM3NjFAMTQ2MTY3MDM0MjU5OQ%3D%3D&el=1_x_7https://www.researchgate.net/institution/Second_University_of_Naples?enrichId=rgreq-6f6baea8-4e64-4b77-a198-380e0368005a&enrichSource=Y292ZXJQYWdlOzMwMTU5MjExNDtBUzozNTUwODUzMDY2MTM3NjFAMTQ2MTY3MDM0MjU5OQ%3D%3D&el=1_x_6https://www.researchgate.net/profile/Luigi_Trojano?enrichId=rgreq-6f6baea8-4e64-4b77-a198-380e0368005a&enrichSource=Y292ZXJQYWdlOzMwMTU5MjExNDtBUzozNTUwODUzMDY2MTM3NjFAMTQ2MTY3MDM0MjU5OQ%3D%3D&el=1_x_5https://www.researchgate.net/profile/Luigi_Trojano?enrichId=rgreq-6f6baea8-4e64-4b77-a198-380e0368005a&enrichSource=Y292ZXJQYWdlOzMwMTU5MjExNDtBUzozNTUwODUzMDY2MTM3NjFAMTQ2MTY3MDM0MjU5OQ%3D%3D&el=1_x_4https://www.researchgate.net/?enrichId=rgreq-6f6baea8-4e64-4b77-a198-380e0368005a&enrichSource=Y292ZXJQYWdlOzMwMTU5MjExNDtBUzozNTUwODUzMDY2MTM3NjFAMTQ2MTY3MDM0MjU5OQ%3D%3D&el=1_x_1https://www.researchgate.net/publication/301592114_Drawing_Disorders_in_Alzheimer%27s_Disease_and_Other_Forms_of_Dementia?enrichId=rgreq-6f6baea8-4e64-4b77-a198-380e0368005a&enrichSource=Y292ZXJQYWdlOzMwMTU5MjExNDtBUzozNTUwODUzMDY2MTM3NjFAMTQ2MTY3MDM0MjU5OQ%3D%3D&el=1_x_3https://www.researchgate.net/publication/301592114_Drawing_Disorders_in_Alzheimer%27s_Disease_and_Other_Forms_of_Dementia?enrichId=rgreq-6f6baea8-4e64-4b77-a198-380e0368005a&enrichSource=Y292ZXJQYWdlOzMwMTU5MjExNDtBUzozNTUwODUzMDY2MTM3NjFAMTQ2MTY3MDM0MjU5OQ%3D%3D&el=1_x_2


2/23

Journal of Alzheimers Disease xx (20xx) xxxDOI 10.3233/JAD-160009IOS Press

1

Review1

Drawing Disorders in Alzheimers Diseaseand Other Forms of Dementia

2

3

Luigi Trojanoa,b and Guido Gainottic,d4aDepartment of Psychology, Second University of Naples, Italy5bS. Maugeri Foundation, Scientific Institute of Telese Terme (BN), Italy6cCenter for Neuropsychological Research, Institute of Neurology, Catholic University, Rome, Italy7dIRCCS Fondazione Santa Lucia, Department of Clinical and Behavioral Neurology, Rome, Italy8

Handling Associate Editor: Jason Brandt9

Accepted 16 March 2016

Abstract. Drawing is a multicomponential process that can be impaired by many kinds of brain lesions. Drawing disorders

are very common in Alzheimers disease and other forms of dementia, and can provide clinical information for the distinction

of the different dementing diseases. In our review we started from an overview of the neural and cognitive bases of drawing,

and from a recollection of the drawing tasks more frequently used for assessing individuals with dementia. Then, we analyzed

drawing disorders in dementia, paying special attention to those observed in Alzheimers disease, from the prodromal stages

of the amnesic mild cognitive impairment to the stages of full-blown dementia, both in the sporadic forms with late onsetin the entorhino-hippocampal structures and in those with early onset in the posterior neocortical structures. We reviewed

the drawing features that could differentiate Alzheimers disease from vascular dementia and from the most frequent forms

of degenerative dementia, namely frontotemporal dementia and Lewy body disease. Finally, we examined some peculiar

aspects of drawing disorders in dementia, such as perseverations, rotations, and closing-in. We argue that a careful analysis

of drawing errors helps to differentiate the different forms of dementia more than overall accuracy in drawing.

10

11

12

13

14

15

16

17

18

19

20

Keywords: Alzheimers disease, constructional apraxia, drawing disorders, frontotemporal dementia, Lewy body disease,

vascular dementia

21

22

INTRODUCTION23

Different kinds of drawing disorders can be found24

in Alzheimers disease (AD) and other forms of25

dementia. Investigations conducted in patients with26

focal brain lesions andwith differentforms of demen-27

tia have shown that many of these disorders can28

be traced back to visual-spatial [15] or planning29

disturbances [610]. These disorders are the main30

determinants of constructional apraxia (CA) and are

Correspondence to: Luigi Trojano, Dept. of Psychology, Sec-

ond University of Naples, Viale Ellittico 31, 81100 Caserta, Italy.Tel./Fax: +39 0823 274784; E-mail:[email protected].

subsumed by lesions affecting the parietal [1114] 31

and frontal regions of the brain [15]. They are usu- 32

ally observed in copying tasks, whereas other forms 33

of drawing disorders, mainly due to disruption of (or 34

impaired access to) the pictorial representations of 35

objects, can be observed on tasks of drawing from 36

memory [1619]. The relationships between clini- 37

cal forms of dementia and the corresponding patterns 38

of drawing disabilities stem, therefore, from the fact 39

that (at least in the early stages of the disease) the 40

brain pathology affects different brain networks pref- 41

erentially. Furthermore, the need for simple tasks 42with rich informational content in the assessment of 43

ISSN 1387-2877/16/$35.00 2016 IOS Press and the authors. All rights reserved
mailto:[email protected]:[email protected]


3/23

2 L. Trojano and G. Gainotti / Drawing Disorders in Dementia

dementia prompted the implementation of drawing44

tasks, such as the clock drawing test (CDT) [2022],45

in place of those classically used for assessing CA in46

patients with focal brain lesions. The present review47

will therefore include several sections. The first will48

illustrate the neural and cognitive bases of draw-49

ing. The second will focus on the drawing tasks50

more frequently used in dementia. In the third sec-51

tion we will pass to analyze drawing disorders in52

AD, from the prodromal stages of the amnesic mild53

cognitive impairment (aMCI) to the stages of full-54

blown dementia, both in the sporadic forms with late55

onset in the entorhino-hippocampal structures and in56

those with early onset in the posterior neocortical57

structures. In thenextsection we will consider thefea-58

tures of drawing disorders that could differentiate AD59

from vascular dementia and from the most frequent60forms of degenerative dementia, namely frontotem-61

poral dementia (FTD) and the Lewy body disease62

(LBD). In the last section we will take into account63

some peculiar aspects of drawing disorders in demen-64

tia, paying particular attention to the closing-in65

phenomenon [2325].66

Drawing disorders in dementia are a field that is67

attracting an ever-growing interest. A 2005-2015 lit-68

erature search limited to Pubmed database and to69

papers in English language, with the terms [drawing70

OR copying OR constructional] AND dementia71

as keywords, identified 793 papers. Faced with this72

huge amount of studies, we realized that our work73

could not be exhaustive but wanted to provide a74

comprehensive review including the most relevant75

clinical aspects of drawing disorders.76

NEURAL AND COGNITIVE BASES OF77

DRAWING78

Poppelreuter [26] observed that some brain dam-79

aged patients may be impaired in a series of activities80

requiring a careful control of vision on action, such81

as drawing, but Kleist [27] was the first to propose the82

term CA to designate a specific disturbance which83

appears in formative activities (such as assembling,84

building, or drawing) in which the spatial form of85

the task is missed, although there is no apraxia of86

the single movements. In subsequent years, drawing87

became progressively more popular in neuropsycho-88

logical assessment, and a lot of studies have been89

aimed at comprehending its neural and cognitive90

bases in brain-damaged patients and in healthy sub-91jects [22, 28].92

Studies on focal brain-damaged patients have often 93

provided contrasting evidence about the lateralization 94

and the intra-hemispheric locus of lesions provoking 95

drawing disabilities. Since 1962 Benton [29] under- 96

lined that the subjective nature of the clinical method, 97

and the large variability of the tasks used to study 98

constructive disorders could account for the incon- 99

sistencies found in previous literature. Benton thus 100

suggested using gradedtests with precise scoring pro- 101

cedures to improve the comparability of results [29]. 102

The ensuing authors adopted this pivotal standpoint 103

in trying to comprehend the neural basis of drawing 104

through the analysis of performance in focal brain 105

damaged patients. However, many studies did not 106

detect differencesin prevalenceor severityof drawing 107

disorders after right or left hemispheric lesions (e.g., 108

[2, 6, 3032]). Then, prompted by a first systematic 109survey about the relationships between visuocon- 110

structive disabilities and hemispheric locus of lesion 111

[12], several authors attempted to ascertain whether 112

qualitative differences existed between the drawing 113

disorders resulting from right or left hemispheric 114

lesions. On one hand, correlational studies demon- 115

strated that drawing disabilities are tightly related 116

with scores on visual-perceptual tasks in right but 117

not in left brain-damaged patients, e.g., [3, 4, 18, 33]. 118

On the other hand, drawing disorders in left-brain 119

damaged patients have been ascribed to a planning 120

disorder [6, 7], but this interpretation has not been 121

supported by several empirical studies [8, 9, 17, 18, 122

33]. In one of the most recent systematic attempts at 123

identifying the neural structures involved in copy- 124

ing a complex figure, in a wide sample of focal 125

brain-damaged patients, voxel-based morphometry 126

revealed that different lesions in the two hemispheres 127

were significantly correlated with different aspects of 128

the drawing production [34]. 129

Rather inconclusive results on the neural bases of 130

drawinginhealthyindividualshavealsobeenobtained 131

by functional neuroimaging studies, because the lat- 132

ter have been constrained by the relevant artifacts 133

induced by the hand and arm movements, and have 134

oftenassesseddrawing-relatedtasksratherthanactual 135

drawing [14, 3538]. Only very recently the devel- 136

opment of functional MRI-compatible graphic tablets 137

[39] allowed the analysis of hand movements during 138

drawing, and one recent study using such a device 139

reported activation within a wide network extend- 140

ing from the temporo-occipital to parietal and lateral 141

frontal areas bilaterally in healthy participants while 142

they were drawing faces or abstract patterns with dif- 143ferent levels of visual details [40]. 144


4/23

L. Trojano and G. Gainotti / Drawing Disorders in Dementia 3

Neuropsychological and functional neuroimaging145

data are thus consistent with the view that the process146

of drawing involves different cognitive components.147

Several authors tried to single out these components148

within comprehensive models [4143]. All these149

models share the idea that visuospatial processes,150

dedicated planning abilities, and general control pro-151

cesses are involved in drawing [44], although the152

models differ from each other in terms of formal153

characteristics, depth of analysis and sometheoretical154

aspects. It is worth mentioning that Grossi [43], draw-155

ing on the distinction between lexical and sub-lexical156

components of language, proposed the existence of157

two copying procedures: a lexical route and a line-158

by-line procedure. The former would consist of159

motor subroutines, which could be considered as part160

of a constructional lexicon, could be used to draw161well-known figures (such as a square or a face), and162

woulddevelop as a resultof formaleducation andper-163

sonal aptitudes [4546]. The latter procedure would164

not rely on previously acquired constructional repre-165

sentations, could be used for copying novel stimuli,166

and would be only based on a piecemeal spatial anal-167

ysis. Both procedures could be adopted for copying168

complex pictures, but some patients might be con-169

strained to use either one or the other. For instance,170

patients with visual agnosia tend to adopt a slavish171

line-by-line copying procedure for familiar objects172

they cannot recognize [4748]. On the contrary some173

demented patients can draw simple figures success-174

fully but fail at integrating correctly shaped simple175

elements in a coherent whole, because of planning or176

visuospatial defects.177

DRAWING TASKS FREQUENTLY USED IN178

DEMENTIA179

Drawing tasks have been widely used in patients180

with dementia, but drawing performance may vary181

greatly, as a function of the task (copying and free182

drawing cannot be considered equivalent), of the183

stimuli used in the task (more complex stimuli pose184

greater load on visuospatial and planning functions),185

and even of the patients pre-morbid abilities (largely186

dependent on age,educational level, and even cultural187

background) [49].188

Free drawing, in which the patient is required to189

draw a named object (e.g., a clock, a face, an ani-190

mal, or a tool), can reveal information about the191

patients ability to draw complete shapes or a ten-192dency to omit parts and about their ability to organize193

the figure as a whole, with its component elements in 194

the correct spatial relationships. However, this task 195

also relies on non-constructional cognitive abilities, 196

particularly on lexical-semantic knowledge, pictorial 197

representations, and mental imagery [16, 50, 51]. 198

A clear instance of the complexity of free draw- 199

ing tasks is CDT which is widely used in the first 200

steps of detecting cognitive impairment and dementia 201

(Fig. 1), as a stand-alone neuropsychological test or 202

Fig. 1. Drawing a clock face on an empty circle, and putting thehands at 2 : 45. Spatial distortions and errors reflecting impairedsemantic knowledge in drawings by patients with Alzheimers

disease (first row). Planning errors, perseverations, and stimulus-bound responses in drawings by patients with severe forms of

vascular dementia (second row) and of behavioral vascular demen-tia (second row) or behavioral variant of frontotemporal dementia(third row). Gross spatial distortions and perseverations in draw-

ings by patients with Lewy body disease (bottom row). Oftenpatients with dementia are not able to draw the clock hands at

the given time, and some write numbers instead of drawings thehands.


5/23


included into composite screening batteries [20, 21,203

52]. In its classical form [53], CDT requires patients204

to draw the face of a clock with all the numbers205

and set the two hands to 10 after 11. Many admin-206

istration procedures and scoring systems have been207

proposed [5457], but there is no consensus as to208

which is the most useful for dementia screening [20,209

54]. The scoring systems share commonalities, but210

each may reflect different cognitive components and211

be correlated with different areas of brain atrophy212

[58]. However, independently from the scoring pro-213

cedure, performance on CDT in demented patients214

seems to be inversely correlated with lesion load in215

medial temporal lobe, in subcortical structures and in216

periventricular white matter [59, 60] and might pro-217

vide prognostic information for cognitive decline [61,218

62]. Some authors suggested that qualitative analysis219of errors in clock drawing might improve diagnos-220

tic accuracy for dementia screening. For instance,221

errors such as inaccurate time setting, missing hands222

or numbers, and number substitutions or repetitions223

might be particularly useful for dementia screening224

[63]. Using appropriate error classification criteria,225

error analysis in CDT might be useful to differen-226

tiate different types of dementia, since conceptual227

errors are quite frequent in AD [64], but in a recent228

longitudinal study conceptual errors and persevera-229

tions were more frequently observed in the advanced230

stages of all dementing diseases, whereas spatial231

and planning errors were more frequent in mild-to-232

moderate dementia [65]. Errors in setting numbers233

and clock hands seem to correlate with regional234

hypometabolism in bilateral parietal and posterior235

temporal areas and in the right middle frontal gyrus236

[58]. By the same token, it has been recently observed237

that different types of errors in demented patients238

performance on CDT might correlate with atrophy239

in different brain regions, within frontal, parietal and240

temporal lobes [66].241

One strategy for improving diagnostic accuracy of242

CDT is to compare the drawing to command con-243

dition with a copy condition, in which subjects are244

required to copy a pre-drawn clock, since in the copy245

condition performance would rely only on a subset of246

cognitive functions centered on visuoperceptual and247

constructional skills [16]. However, although copy-248

ing seems to directly assess the patients ability to249

reproduce a figure, even this task can imply prob-250

lem solving and executive abilities and is affected by251

age, educational level, and even cultural background252

[49]. In assessing copying abilities, it is therefore253necessary to adopt standardized tasks, with solid254

Fig. 2. Copying of geometrical drawings: cube (modelon thetop).

Gross simplifications, spatial distortions, errors of perspective inpatients withAlzheimers disease (firstrow) and vascular dementia(second row). Simplifications, distortions and perseverations in

patients with moderate to severe frontotemporal dementia (thirdrow). The bottom figures show instances of near closing-in (on the

left) and of adherent closing-in (on the right).

normative data, and to use different kinds of stim- 255

uli, well-known or novel, of graded complexity from 256

simpleshapes, such as circles andsquares, to complex 257

figures, such as a cube (Fig. 2), two intersecting pen- 258

tagons (Fig.3), or the Rey-Osterreith Complex Figure 259

(ROCF) [67, 68]. In particular, the copy of the ROCF 260

(Fig. 4) is often used in the assessment of drawing 261

abilities in individuals with dementia, and provides 262

an opportunity to assess copying procedures together 263

with copying accuracy [67]. The copy of the ROCF 264

has high sensitivity for detecting brain damage, since 265it involves a wide network of brain areas, including 266


6/23


Fig. 3. Copying of geometrical drawings: interlocking pentagons (model on the top). Simplifications and distortions in drawings by patientswith Alzheimers disease (first row); perseverations, simplifications and spatial distortions in patients with vascular dementia (second row);

simplification, distortions and planning errors in patients with moderate to severe frontotemporal dementia (third row). The bottom figuresshow instances of near closing-in (on the left) and of adherent closing-in (on the right).

frontal, superior temporal, posterior parietal and mid-267

dle occipital cortex, more extensively in the right268

hemisphere [69]. Several neurofunctional studies in269

dementia showed that accuracy in copying the ROCF270

is related to metabolic rate in bilateral temporal-271

parietal cortex and occipital lobe, and in right frontal272

lobe, whereas the procedure used to copy, namely the273

tendency to draw first the main organizational lines of274

the model, correlated with metabolism in right lateral275

temporal cortex [70]. It is worth mentioning, how-276

ever, that neurofunctional studies comparing copying277

of different kinds of stimuli in dementia demon-278

strated that the neural correlates of copying tasks, 279

mainly centered in the posterior brain regions, differ 280

as a function of task complexity [71]. Moreover, it 281

appears advisable to complement evaluation of accu- 282

racy in copying tasks with a qualitative analysis of 283

the patients production, which can provide further 284

relevant information. This fact has been underlined 285

above with respect to the CDT, and is also sug- 286

gested by recent studies showing that the copy of 287

two intersecting pentagons (Fig. 3), included in the 288

Mini-Mental State Examination [72], may provide 289

elements for differential diagnosis of degenerative 290


7/23


Fig. 4. Copying of Rey complex figure (model on the top). Gross spatial distortions and simplifications in patients with Alzheimers disease(firstrow); planning errors, spatial distortions, and perseveration in patientswith vascular dementia(second row)and frontotemporaldementia(lowest row). In the left sided column are reported drawings of patients with mild forms and in the right-sided columns those of patients

with severe forms of dementia. Note in the right top figure that the reproduction of the single subcomponents is relatively spared, but spatialrelationships among them are lost; in the right bottom figure an instance of rotation.

dementias [73, 74]. Although error analysis in draw-291

ing tasks may not reveal a straightforward procedure292

[22, 28, 44], several clinical phenomena observed in293

drawing have been sufficiently characterized in their294possible cognitive and neural correlates, as it will be295

discussed in the next sections of this paper.296

It is important to stress, however, that in the diag-297

nostic work-up for dementia performance on drawing298

tasks, and qualitative error analysis should be com-299

plemented by assessment of the deficits that can300

contribute to constructional apraxia. As it will be301

discussed later, mechanisms inducing errors in spon-302

taneous drawing or in copying may encompass basic303

deficits in visuoperceptual abilities, spatial attention,304

spatialworkingmemoryorspatialplanning,andexec-305

utive function [75]. Particularly crucial in this respect306

seem to be visuospatial perception whose different 307

aspectscanbeimpairedindementia[44,7578].Visu- 308

ospatial perception is often assessed by tests such as 309

Bentons Judgement of Line Orientation, in which 310subjects are required to identify the lines that have 311

the same angulation as those presented as stimuli 312

[79], or by comprehensive test batteriesassessing sev- 313

eral aspects of visuoperceptual processing, such as 314

Warrington and James Visual Object and Space Per- 315

ception Battery [80], or the Battery for Visuospatial 316

Abilities[4,81].Themechanismsunderlyingdrawing 317

disorders can also be clarified by means of plan- 318

ning/executive tests, such as Trail Making test [82] or 319

tests for figural fluency [83], and other tests that have 320

been used to assess patients with focal lesions. Motor 321

programming deficits have also to be considered, 322


8/23


because patients with dementia may be affected by323

limb apraxia or optic ataxia with a strong impact on324

drawing.Therefore,thesetypesoftestshavetobeused325

to evaluate the different forms and the different stages326

of dementia. To establish an algorithm for assessing327

the different abilities involved in the drawing pro-328

cess and the disorders than can cause constructional329

apraxia might provide a powerful tool to understand330

the genesis of drawing disorders and of the different331

types of errors in dementia.332

DRAWING DISORDERS IN ALZHEIMERS333

DISEASE334

The observation that many brain regions in both335

hemispheres are involvedin different aspects of draw-336

ing, and by implication of constructional tasks, can337

provide an explanation of the classical findings sug-338

gesting that CA is related to poor intellectual abilities339

in focal brain damaged patients [30], and can repre-340

sent an index for diffuse cognitive deterioration, after341

lesions in either left [84] or right hemisphere [85].342

According to old studies, in AD drawing disorders343

are present since early stages of the disease, and their344

severity increases as the illness progresses [22, 28].345

However, more recent investigations have shown that346

in the early stages of AD simple models can be eas-347

ily copied and that drawing disabilities are observed348

only with rather complex tasks [86, 87].349

Several variables must therefore be taken into350

account in the study of drawing disorders in AD.351

Among these variables we will separately consider:352

(1) the drawing tasks used in AD and the corre-353

sponding kinds of errors; (2) the drawing disorders354

in prodromic (aMCI) and clinical forms of AD; (3)355

the drawing disabilities in early and late onset forms356

of AD; (4) the drawing disorders in different stages357

of evolution of AD; (5) the mechanisms giving rise358

to drawing disorders in AD patients.359

The drawing tasks used in AD and the360

corresponding kinds of errors361

An attempt at a systematic description of AD362

patients errors in a free drawing task [9] has shown363

frequent occurrence of simplifications, spatial alter-364

ations, and lack of perspective. However, since free365

drawing poses a heavy load on semantic memory,366

errors on this task (e.g., simplifications in drawing a367

house) may derive from impaired access to semantic368

knowledge or to impaired visuoperceptual process-369ing [51]. For this reason, spontaneous drawing could370

be impaired in the early stages of AD, while copy- 371

ing may deteriorate later [88]. The reproduction of 372

complex figures is particularly sensitive to the pro- 373

gression of the disease [89]. For instance, the ROCF 374

may be reproduced in a simplified way, with single 375

constitutive elements put one after the other, even in 376

early AD (see Fig. 4). In these cases, patients seem 377

able to recognize and reproduce single well-known 378

elements but are unable to reproduce complex spa- 379

tial relationships correctly. Another simplification 380

error may consist in the reproduction of more familiar 381

or simpler figures instead of more complex ones (e.g., 382

a square instead of a diamond). As the disease pro- 383

gresses, patients usually become unable even to draw 384

simple figures correctly, as they no longer had access 385

to well-consolidated motor subroutines and can show 386

peculiar patterns of behavior, such as the closing-in 387phenomenon [2325, 90, 91] (see Figs. 2 and 3). 388

Drawing disorders in prodromic (aMCI) and 389

clinical forms of AD 390

Drawing disorders are not usually present in MCI, 391

at least when copying of simple figures is employed 392

[92]. However, a different pattern might emerge as a 393

function of clinical characteristics of MCI patients, 394

particularly if more complex tasks are used. In a 395

study on brain morphometric correlates of CDT 396

in patients with MCI or AD, Thomann et al. [86] 397

observed that, even though MCI patients did not dif- 398

fer from matched healthy controls in copying simple 399

geometrical drawings, CDT could discriminate MCI 400

from matched healthy controls and AD from MCI 401

patients. Analogously, in a further brain morpho- 402

metric study on MCI and AD patients, accuracy in 403

copying the ROCF was significantly different in MCI 404

patients, AD patients, and healthy controls [87]. It 405

has also been observed that MCI patients make con- 406

ceptual and graphic errors more often than matched 407

healthy controls, and that AD patients score lower 408

and make significantly more conceptual, graphic, and 409

spatial-planning mistakes than MCI individuals, thus 410

suggesting that a detailed scoring system is necessary 411

to differentiate individuals with MCI from healthy 412

adults [63, 93]. Scoring systems focusing on hand 413

and number placement might better differentiate MCI 414

individuals from healthy controls, but the solidity of 415

CDT as a screening tool for MCI has been ques- 416

tioned [94]. In this respect, it should be taken into 417

account that cognitive characteristics of MCI individ- 418

uals might correlate with performance on CDT. It is 419very important to acknowledge at this point that MCI 420


9/23


patients are rather heterogeneous from the etiological421

point of view andthat amnesic MCI canbe considered422

as that most related to development of AD [95, 96].423

In line with this caveat is the fact that, a recent study424

contrasting MCI of the amnesic, dysexecutive, or425

multi-domain type on CDT [97] confirmed that aMCI426

did not differ from a matched sample of healthy con-427

trols, whereas both dysexecutive and multi-domain428

MCI individuals made a significantly larger number429

of errors in hand and number placement.430

Poor performance on CDT might predict future431

development of AD in patients with MCI [98], but432

not all qualitative scoring procedures might be suf-433

ficiently sensitive to capture longitudinal changes434

of CDT from MCI to AD [61]. A recent meta-435

analysis [99] confirmed that the CDT might reveal436

a useful measure of cognitive decline over time, and437that conceptual clock drawing errors (with particu-438

lar mention of hand and number placement) would439

be most informative about the cognitive decline.440

This meta-analysis also showed that CDT might441

differentiate individuals who are going to develop442

dementia in subsequent years, thus being included443

among the tests helpful for predicting conversion to444

dementia.445

Drawing disabilities in early and late onset446

forms of AD447

The structural or functional neuroimaging abnor-448

malities correlate with the clinical features in AD449

patients, because typical late-onset patients (LOAD)450

present with memory disorders and medial tempo-451

ral lobe atrophy, whereas focal neocortical forms452

of AD prevail in early-onset patients (EOAD) and453

are associated with executive-behavioral disorders,454

logopenic aphasia, and (in posterior cortical atrophy)455

with progressive visual-spatial disorders [100, 101].456

Some authors have, therefore, examined the relations457

between age at onset and drawing disorders in AD458

patients, and reported a high prevalence of visuocon-459

structional disabilities, assessed with the copy of the460

ROCF, in EOAD patients [101103]. In some stud-461

ies [102, 103], but not in others [101], these drawing462

disorders were associated with visuospatial deficits.463

Furthermore, some studies [104] have shown that464

the presence and severity of drawing disorders cer-465

tainly depends on the AD phenotype and not just on466

age. Most frequent and severe drawing disorders are467

found in posterior cortical atrophy (PCA), which is468

associated with prominent visuospatial impairments469and relative preservation of memory, insight, and470

judgment [105]. PCA is associated with atrophy in 471

the occipital, parietal, and posterior temporal lobes 472

[106], but recently a lesion of long white-matter 473

tracts, including the superior and inferior longitudinal 474

fasciculi and the inferior fronto-occipital fasciculus, 475

has also been documented [107]. This early-onset 476

AD-related pathologygives rise to complex visuospa- 477

tial difficulties, such as visuoperceptual impairments 478

[108], optic ataxia, simultanagnosia, gaze apraxia 479

(often with the complete spectrum of Balint syn- 480

drome), and even egocentric [109, 110] or allocentric 481

[111] unilateral spatial neglect. All such visuospa- 482

tial impairments have a strong impact on drawing 483

abilities, affecting both spontaneous drawing and 484

copying, and giving rise to severe displacements of 485

drawing elements, omissions, gross spatial distor- 486

tions (exploded drawings) [112]. 487

Drawing disorders in different stages of 488

evolution of AD 489

A strictrelation between the progression of visuop- 490

erceptual spatial impairment and the progressive 491

deterioration of the performance on copying the 492

ROCF has been reported [89] (see Fig. 4). Mild AD 493

usually spares basic visual sensory processes and 494

affects all high-order visual processes, along both the 495

ventral (occipito-temporal) and the dorsal (occipito- 496

parietal) visual streams [113]. Caine and Hodges 497

[114] carried out two separated studies to investigate 498

ventral and dorsal visual functions in a sample of 499

AD patients at early stages of the disease. A small 500

subsample of patients were impaired on visuospatial 501

tasks, thus suggesting that this small group of patients 502

with prominent visuospatial disorders might repre- 503

sent one end of a continuous spectrum at the other 504

end of which are patients affected by a focal degener- 505

ative dementia involving occipito-parietal cortex, the 506

so-called PCA [115118]. In the moderate stages of 507

AD, visuospatial impairments would become more 508

evident, and play a relevant role in the develop- 509

ment of constructional disorders [89]. More recently, 510

Guerin et al. [119] examined the cognitive mech- 511

anisms underlying the constructional performance 512

of AD patients at different stages of the disease by 513

means of a copying task and of visuospatial tasks 514

measuring spatial exploration (visual search tasks), 515

judgment of spatial relations, and planning abilities. 516

The results suggested that cognitive defects underly- 517

ing constructional impairment in patients with AD 518

involved the early phases of spatial-constructional 519processing, likely exploration and judgment of 520


10/23


spatial relationships, rather than the late stage of521

planning.522

Mechanisms giving rise to drawing disorders in523

AD patients524

The drawing impairment in AD patients may stem525

from different cognitive mechanisms.526

In recent years, the different mechanisms giving527

rise to CA in AD patients have been tackled with by528

means of morphometric and neurofunctional studies.529

However, such studies did not provide strongly con-530

vergent data. For instance, in a sample of AD and531

MCI patients overall score on CDT was inversely532

correlated with atrophy in the middle and superior533

temporal gyri (BA 21 and 22) bilaterally, but more534

strongly in the left hemisphere, and in the left entorhi-535nal area (BA 28) [86], whereas in a sample of AD536

patients impaired performance on the ROCF copying537

task was correlated with atrophy in several fronto-538

temporo-parietal regionsin the right hemisphere [87].539

In a further study [5], impaired performance on a540

copying test was correlated with bilateral parietal541

atrophy. The differences between such studies might542

originate from different characteristic of the AD sam-543

ples, and from the neuropsychological measure used544

for assessing drawing. Severalrecent neurofunctional545

studies are illustrative of these sources of variabil-546

ity. In a large sample of AD patients, Shon et al.547

[120] found that drawing from memory or copy-548

ing a clock face correlated with regional glucose549

metabolism in bilateral temporo-parietal regions, but550

these correlations changed as a function of dementia551

severity. Matsuoka et al. [58] reported another rep-552

resentative instance of variability in neurofunctional553

findings depending on the neuropsychological mea-554

sure in a large sample of drug-nave AD patients,555

in whom overall performance on CDT, the partial556

score on locating numbers of the clock and the557

partial score on setting the hands were correlated558

with atrophy in distinct parietal, posterior tempo-559

ral and frontal regions. A PET study using copying560

of simple geometrical figure as a measure of con-561

structional abilities showed, instead, an exclusive562

correlation of copying performance with the left infe-563

rior parietal lobule and the left inferior frontal gyrus564

[121].565

These data explain why the diffuse involvement of566

several areas of the brain in AD can impair draw-567

ing performance since the early clinical stages of568

the diseases, particularly when drawing is assessed569by means of complex stimuli. Therefore, drawing is570

characteristically affected in AD, as well as visuospa- 571

tial cognition in general [76]. 572

DRAWING DISORDERS IN THE 573

DIFFERENTIAL DIAGNOSIS BETWEEN 574

AD AND VASCULAR DEMENTIA OR 575

OTHER DEGENERATIVE FORMS OF 576

DEMENTIA 577

The widespread involvement of several neural net- 578

works in drawing can explain the clinical observation 579

of frequent drawing disorders in AD, but also implies 580

that drawing disorders are often found in other forms 581

of dementia, bothof vascularand degenerative nature. 582

In recentyearsa growingnumber of studies attempted 583

to ascertain whether drawing could provide addi- 584

tional information for refining clinical differentiation 585

among dementias. Here we will provide an overview 586

of such studies, from which it will emerge that it can 587

be difficult to disentangle drawing disorders in AD 588

from those shown by patients affected by other kinds 589

of progressive dementias, and that some distinctive 590

features might be captured by qualitative analysis 591

of drawing productions (see Table 1). As it will be 592

argued in the next section, it is possible, however, 593

that if patients with different etiologies share some 594

neural lesions or cognitive defect, they will produce 595

similar error types in drawing. 596

Drawing disorders in vascular dementia 597

Since AD and vascular dementia (VaD) are the 598

most common forms of dementia in old age, several 599

investigations tried to evaluate if drawing disor- 600

ders can be useful to differentiate the two forms 601

of dementia. In particular, a few studies compared 602

drawing abilities in AD and VaD patients on the 603

CDT, without providing convergent evidence. Looi 604

and Sachdev [122] reviewed 11 studies comparing 605

the performance of VaD and AD patients in the area 606

of constructional apraxia, using tests such as block 607

design, CDT, or the ROCF copying task. Eight of 608

the studies showed no difference between the two 609

groups, while in three it was found that the AD 610

were less impaired. Wiechmann et al. [123] did not 611

find difference in accuracy on CDT between AD 612

and VaD. Instead, in a study on a sample of outpa- 613

tients matched for general cognitive impairment VaD 614

scored significantly lower than AD on CDT, likely 615

related to poorer executive abilities in VaD patients 616

[124]. Kitabayashi et al. [125] stratified their patient 617sample by general cognitive impairment, and found 618


11/23


Table 1

Summary of studies comparing drawing disorders in Alzheimers Disease and in other forms of dementia

First author Drawing test Findings

Comparison AD-VaD

Wiechmann [123] CDT No difference in accuracy

Kitabayashi [125] CDT Conceptual and spatial/planning errors in AD in all disease stages;conceptual and graphic errors in moderate-severe VaDFukui [126] CDT, figure copying No difference at mild disease stages; sVaD worse

than AD on both tasks at severe stagesGraham [127] ROCF No difference between AD and sVaDComparison AD-FTD

Rascovsky [131] CDT, figure copying AD worse than FTD on overall CDT score; no differences in figure copyingBlair [132] CDT AD worse than FTD; less conceptual, spatial

and planning errors in FTD than in ADRazani [133] CDT AD worse than right- and left-onset FTD

Thompson [134] Figure copying AD worse than FTD; less spatial errors,but more perseverations in FTD than in AD

Possin [135] Simplified ROCF AD worse than FTD

Perri [136] ROCF AD worse than FTDGorno-Tempini [137] ROCF AD worse than non fluent PA

Grossi [138] ROCF No difference between AD and bvFTDGasparini [139] ROCF No overall difference between AD and bvFTD;

AD better than bvFTD in drawing detailsRascovsky [141] Figure copying Steeper decline of copying abilities in FTD with respect to ADMendez [142] Figure copying No 2-year decline in FTDComparison AD-LBD

Nervi [152] Figure copying LBD worse than AD

Crowell [153] Figure copying LBD worse than ADTiraboschi [154] Figure copying LBD worse than AD; no differences in pentagon copyingCahn-Wiener [156] CDT No differences in overall score, but more conceptual and planning errors in LBD

Hamilton [157] Clock copying No differences, but lower scores predicted steeper decline in LBDCagnin [158] CDT LBD worse than AD

Palmqvist [159] CDT, cube copying LBD worse than ADGnanalingham [162] CDT, clock copying No difference in CDT; LBD worse than AD in clock copying

Ala [164] Pentagon copying LBD worse than ADCormack [165] Pentagon copying LBD worse than ADConnor [166] Pentagon copying No difference

Caffarra [167] Figure copying, No difference in figure copying,pentagon copying worse scores in LBD on an analytic scoring system for pentagon copying

AD, Alzheimers disease; VaD, vascular dementia; sVaD, subcortical VaD; FTD, frontotemporal dementia; bvFTD, behavioral variant of

FTD; PA, progressive aphasia; LBD, Lewy body dementia; CDT, Clock drawing test; ROCF, copy of Reys complex figure. Reference

numbering follows the text.

that CDT can well discriminate patients with AD or619

VaD from healthy controls, but the main differences620

between AD and VaD patients were in the relative621

percentage of error types, with AD patients showing622

many conceptual and spatial/planning errors inde-623

pendently from disease severity and VaD patients624

showing conceptual errors and graphic difficulties625

only in presence of moderate cognitive impairment626

(see Fig. 1). Partially different findings have been627

reported by Fukui et al. [126], who did not find dif-628

ferences between AD and subcortical VaD patients629

at mild disease stages on CDT and figure copying,630

whereas at later stages VaD achieved significantly631

lower scores than AD in both drawing tasks. Graham632

et al. [127] also found that patients with subcortical633

VaD were more impaired than AD patients on exec-634

utive/attentional functioning, and visuospatial and635

perceptual skills. In their recent review on CDT, Tan 636

et al. [128] concluded that the overall accuracy on 637

CDT in AD versus VaD patients did not provide 638

consistent results, whereas qualitative analyses might 639

reveal useful differences between the two groups of 640

patients, since VaD patients tended to show more fre- 641

quent executive/planning or perseveration errors than 642

AD patients. 643

Drawing disorders in frontotemporal dementia 644

The relative preservation of visuospatial and 645

constructional abilities is a feature capable of dis- 646

tinguishing FTD from other degenerative dementias 647

and, notably, AD [129, 130]. Consistent with this 648

idea, a few studies observed better performance on 649CDT in FTD compared with AD. For instance, in 650


12/23


a sample of autopsy-verified FTD patients, Rascov-651

sky et al. [131] observed similar scores in copying652

geometrical figures, but higher overall CDT scores653

in FTD compared to AD patients. In the same vein,654

Blair et al. [132] reported higher overall scores, and655

fewer conceptual, spatial, and planning errors in FTD656

compared to the AD group (see Fig. 1). Compar-657

ing the performance of AD patients with those of658

patients with right and left FTD, Razani et al. [133]659

also confirmed that the AD group displayed sig-660

nificant impairments in visual-constructional ability661

relative to the two FTD groups. On the other hand,662

Thompson et al. [134] observed that FTD patients663

outperformed AD patients in copying geometrical664

figures of different complexity showing a lower num-665

ber of spatial errors, but were also characterized by666

a higher number of perseverative errors and of poor667organization of the copy than AD patients. Thus spa-668

tial errors were highly predictive of AD, whereas669

organizational errors, perseverations, or overelabo-670

rated copies were significant predictors of FTD (see671

Figs. 24). Data supporting the hypothesis of a preva-672

lence of spatial errors in the drawings of AD patients673

and of executive disorders in those of FTD patients674

have also been obtained in a study by Possin et al.675

[135], that used quantitative morphometric evalua-676

tion of structural MRI. These authors reported that677

a sample of patients with the behavioral variant of678

FTDperformedbetter than a sample of AD patients in679

copying drawings, but also observed that the drawing680

impairments in the two patient groups might be based681

on different cognitive and neuroanatomical corre-682

lates. In AD patients poor figure copy was associated683

with performance on spatial perception and attention684

tasks, and correlated significantly with volumes in685

the right parietal cortex. In FTD patients, instead,686

performance on figure copy correlated significantly687

with scores on spatial planning and working memory688

tasks, and correlated with right dorsolateral pre-689

frontal cortex volumes. These observations would fit690

clinical data demonstrating that patients with promi-691

nent behavioral disorders and frontal involvement692

(behavioral variant FTD, bvFTD) outperformed AD693

patients on ROCF immediate reproduction [136],694

and that patients with prominent language deficits695

(non-fluent progressive aphasia) did not differ from696

healthy controls on cube copying and ROCF imme-697

diate reproduction [137]. Therefore, several studies698

confirmed the basic tenet that visuospatial skills and699

drawing abilities are relatively preserved in FTD,700

but at least two studies did not detect differences701in overall accuracy on ROCF immediate reproduc-702

tion between bvFTD and AD patients [138, 139]. It 703

is thus possible that during the disease course FTD 704

patients might show relatively preserved construc- 705

tional abilities in the earlystages, whereas differences 706

between FTD and AD might blur in late stages of 707

the disease. This hypothesis might find support in 708

a meta-analysis showing that constructional abilities 709

assessed on copying the ROCF quickly deteriorate 710

as FTD progresses [140], and in a longitudinal study 711

showing a trend toward a steeper decline of copy- 712

ing abilities in FTD with respect to AD [141, but for 713

contrasting findings see 142]. 714

Divergences about drawing disorders in FTD 715

will be reconciled by taking into account the het- 716

erogeneous clinical features and neuropathological 717

underpinnings of the disease [108, 143]. In this 718

respect, it is important to underline a distinction 719between the drawing abilities of patients with bvFTD 720

and those of patients with the temporal variant or 721

semantic dementia (SD), because both are relatively 722

spared on copying tasks, but the latter can be either 723

enhanced or selectively impaired in drawing from 724

memory the typical shape of common objects. The 725

seemingly contradictory data about drawing abil- 726

ities in SD are illustrated by results obtained by 727

Bozeat et al. [19] and by Miller et al. [144146]. 728

Bozeat et al. [19] showed that when SD patients 729

are required to produce drawings of concrete objects 730

from dictation of their names, their drawings are 731

characterized by a loss of the distinctive features of 732

the represented objects. On the other hand, Miller 733

et al. [144146] showed that in the early stages of 734

SD, patients with left anterior temporal lobe atrophy 735

and severe anomia increase their drawing abilities, 736

copying with great precision and sometimes become 737

artists. In these cases, artworks show an enhanced 738

tendency to realism. This improvement of visuocon- 739

structive and artistic abilities was attributed by Miller 740

et al. [144, 146] to the sparing of the parietal lobe 741

and of the right hemisphere, considered as dominant 742

for the three key features of visual artistry: visual 743

constructive ability, spatial attention,and internal rep- 744

resentation. The apparent contradiction between data 745

reported by Bozeat et al. [19] and by Miller et al. 746

[144146] is probably due to the different stage of 747

evolution of SD in which observations were gath- 748

ered (an early stage of SD, with unilateral left sided 749

atrophy in the case of Miller et al. [144146] and 750

a later stages of SD, with bilateral anterior tempo- 751

ral lobe (ATL) atrophy in the case of Bozeat et al. 752

[19] observations). Gainotti [147, 148] has, indeed, 753recently reviewed data showing that the format of 754


13/23


conceptual representations is mainly verbal in the left755

ATL and mainly sensorial/pictorial in the right ATL.756

In the early stages of SD the atrophy of the left ATL757

can release, with a mechanism labeled paradoxical758

functional facilitation by Kapur [149] the pictorial759

representations stored in the right ATL, leading to an760

enhancement of pictorial abilities. On the other hand,761

in later stages the bilateralspread of atrophy provokes762

a complete disruption of verbal and non-verbal con-763

ceptual representations of concrete entities, leading764

to an inability to draw them from memory.765

Drawing disorders in Lewy body dementia766

In LBD, visuospatialdifficulties areoften early and767

prominent [150]. Visuospatial tasks, such as object768

size discrimination, formdiscrimination, overlapping769figure identification, and visual counting tasks, may770

reveal more impaired performances in LBD than in771

AD, and these defects likely contribute to the dispro-772

portionate impairment in constructional tasks in LBD773

patients [151153].774

In line with these observations, case series of775

pathologically verified LBD [154] as well as sys-776

tematic reviews [155] confirmed that an impairment777

in copying geometrical figures is very common in778

LBD patients since early stages of the disease. How-779

ever, several studies employing CDT did not report780

significant differences in overall total score in LBD781

and AD patients [156, 157], likely because of the782

complexity of the task (see Fig. 1), although at least783

two studies reportedsignificantly lower overallscores784

in LBD versus AD patients [158, 159]. A SPECT785

study showed that LBD patients with defective786

performance on CDT presented more marked hypop-787

erfusion in a frontal-subcortical network, involving788

frontal eye fields and the thalamus, with respect to789

LBD patients with normal CDT scores [160]. A790

more recent neurofunctional investigation by PET791

revealed, instead, that, after controlling for overall792

cognitive impairment, there was a direct associa-793

tion between frontoparietal dysfunction and impaired794

CDT performance in LBD [161].795

The early and prominent impairment of visuospa-796

tial processing in LBD led some authors to suggest797

that LBD patients may be characterized by parallel798

impairments in free drawing and in figure copying799

since early stages of the disease whereas AD patients800

would show relative sparing of figure copying ([162],801

but also see [163]). Likely because of its complexity,802

copy of the ROCF has not been extensively used in803LBD, whereas some interesting observations came804

from the copy of the intersecting pentagons included 805

in the MMSE. Ala et al. [164] found that patients with 806

LBD were more likely than those with AD to copy 807

pentagons incorrectly, and suggested that this impair- 808

ment might help identifying LBD patients. Cormack 809

et al. [165] confirmed that LBD patients show signifi- 810

cantly lower performance in copying pentagons with 811

respect to AD; moreover, while in AD the impair- 812

ment in copying figures was correlated with general 813

cognitive deterioration, in LBD drawing was corre- 814

lated with visuospatial tasks only, thus suggesting the 815

existence of a specific defect in this disease. Differ- 816

ent scoring criteria can likely explain the reason why 817

other authors failed to detect differences between 818

LBD and AD in pentagon copying [108, 159], but 819

more extended copying tests might show signifi- 820

cantly lower performance in LBD with respect to 821AD patients [108]. Recently, it has beendemonstrated 822

that qualitative analysis of the pentagon copy test can 823

provide additional information for distinguishing AD 824

from LBD patients, since a lower number of angles 825

is among distinctive features of LBD patients copies 826

[73, 167], and can also predict development of LBD 827

in MCI patients [74]. 828

PECULIAR ASPECTS OF DRAWING 829

DISORDERS IN DEMENTIA 830

In the previous sections of this paper it has been 831

repeatedly underlined that qualitative observations 832

(i.e., errors analysis) might provide additional infor- 833

mation about the cognitive and neural correlates of 834

drawing disorders in patients with different kinds 835

of dementia. Some qualitative phenomena observed 836

in drawing tasks in demented patients deserve, 837

therefore, some brief comments, because of their fre- 838

quency and their potential clinical usefulness. 839

Graphic perseveration 840

Generally speaking, perseveration can be consid- 841

ered among productive (or positive) pathological 842

signs, since it consists in iterative behavioral 843

responses, not adequate to the current stimulus (see 844

Figs. 24) [168]. Patients may produce the same fig- 845

ure repeatedly, in response to only one stimulus, or 846

replicate stimulis elements (continuous persevera- 847

tion, accordingto Sandson andAlbert) [169];on other 848

occasions, patients may inappropriately draw figures 849

already drawn in previous trials, instead of repro- 850

ducing the current stimulus (recurrent perseveration) 851[169]. 852


14/23


AspecifickindofperseverationisobservedinCDT,853

when patients produce repeatedly the same numbers,854

or start numeration over and over (see Fig. 1). These855

errorsinclockdrawingshavebeendescribedinasmall856

proportionofADpatients[170],butwouldincreasein857

moderate to severe stages of the disease [88]. Ryan et858

al. [171] compared graphic performances on the Ben-859

der visuomotor Gestalt test and onCDT in AD with or860

without wandering, and observed a larger number of861

both continuous and recurrent perseverative errors in862

individualsshowingthe wandering phenomenon,thus863

suggesting a relationship between perseveration and864

behavioral control.865

A first study specifically comparing graphic per-866

severations in patients with AD or VaD showed that867

perseverations were more frequent in VaD than in868

AD [172]. However, Cosentino et al. [173] assessed869patients on CDT and on copying clocks and observed870

that perseveration and closing-in (see below) in AD871

and VaD were more frequent in patients with higher872

numberof white matterlesions andwith more marked873

impairment on executive frontal tasks. These findings874

suggested that the executive impairment associated875

withfrontal-subcortical dysfunctioncontributesto the876

genesis of perseveration in clock drawing in demen-877

tia. Recently, in a retrospectivestudyon a large patient878

sample, De Lucia et al. [174] observed that frequency879

of graphic perseverative errors was similar in AD880

and VaD, and that patients with moderate-to-severe881

dementia produced a significantly higher number of882

perseverations than individuals with mild dementia.883

In both groups graphic perseverations were related884

with frontal and visuoconstructional impairments,885

thussupportingtheviewthatfrontal-executivedefects886

canhamperinhibitionof iterativegraphicproductions.887

In FTD, patients perseveration in drawing would888

be often present also in early stages of the disease889

while specific disturbances in reproducing spatial890

relationships would become evident later during the891

course of the disease [175]. In a systematic analysis892

of errors observed in a wide range of neuropsycho-893

logical tests, Thompson et al. [134] reported that894

FTD patients outperformed a group of AD patients895

in drawing tasks, but the most distinctive feature896

between the two groups were perseveration and poor897

overall organization (see Figs. 24). It is worth898

mentioning, however, that other studies focusing on899

qualitative error analysis [138, 139] did not observe900

a higher proportion of graphic perseverations in FTD901

with respect to AD.902

In LBD, several authors suggested that less903efficient executive control might induce higher fre-904

quency of perseveration than in AD [176, 177], 905

but perseverations in drawing tasks have not been 906

assessed specifically. 907

Rotation 908

In copying stimuli, some patients may respect 909

spatial relationships among constituent elements but 910

reproduce a model with general orientation different 911

from the stimulus, e.g., rotating the reproduction by 912

90 or 180 degrees (see Fig. 4). This behavior has been 913

already mentioned in early studies on constructional 914

apraxia[27,178,179]andhasbeenreported,although 915

rarely, in subsequent descriptions of patients with 916

brain lesions. From a consecutive unselected series of 917

240neurologicalpatients,Solmsetal.[180]identified 918

16 patients who reproduced the ROCF with its major 919axis vertically rotated, independently from accuracy 920

in reproducing spatial relationships among inner ele- 921

ments. Seven of these patients had diffuse cerebral 922

involvement, but all remaining cases showed a lesion 923

involving frontal regions. The authors suggested that 924

thisbehaviorcouldreflectthelackofplanningandver- 925

ification abilities of frontal patients [180]. In another 926

case series of patients with AD, rotation of one or 927

both items of the MMSE interlocking pentagons was 928

reported in about 5% of patients [181]. Frequency of 929

rotation of MMSE pentagons would not differ in AD 930

or LBD patients, but it could progressively increase 931

during the disease course in LBD patients [73]. 932

Recent findings suggest that complex visuopercep- 933

tual and planning mechanisms might contribute to the 934

genesis of rotations. In a retrospective study on a large 935

sample of patients with MCI or degenerative demen- 936

tia who underwent copy and recall of the ROCF, 937

Isella et al. [182] observed rotation on the copy con- 938

dition in 2.7% patients and on recall in 3.3% patients. 939

In a subsequent prospective study on a mixed sam- 940

ple of patients with degenerative dementias or focal 941

brain lesions, rotation at the copy of the ROCF was 942

associated with visuospatial and selective attention 943

impairments, and with more severe temporo-parieto- 944

occipital atrophy or hypometabolism, whereas no 945

specific profile of cognitive impairment distinguished 946

patients with rotation at recall of the ROCF, in whom 947

frontal abnormalities were more frequent [183]. 948

Closing-in 949

In copying tasks, demented patients often show the 950

tendency either to put the pencil directly over the 951model, producing a scrawl, or to overlap the lines 952


15/23


of the model with those of the copy or to draw very953

near to the models elements (see Figs. 2 and 3).954

Such behavior, termed closing-in (CI), has beenfirst955

described by Mayer-Gross [23], and later reported956

more often in demented patients than after focal brain957

lesions [24, 90]. Among dementing syndromes, CI is958

predominant in AD [88, 184], suggesting the idea959

that CI might represent a neuropsychological marker960

for clinical diagnosis of AD [90, 185, 186]. One961

more recent study [25] confirmed that CI is more fre-962

quent in AD with respect to VaD patients, and that963

the two patient groups also tended to show qualita-964

tively different CI phenomena, as AD patients more965

often overlapped at least one element of the copy onto966

the model (adherent-CI), whereas VaD patients more967

often drew close to the model without overlapping968

onto it (near-CI). On the other hand, one retrospec-969tive study showed that CI is as frequent in AD as in970

FTD patients, but did not provide qualitative obser-971

vations [187]. Indeed, in a sample of patients with the972

behavioral variant of FTD, near-CI has been observed973

more frequently than adherent-CI, and was often (but974

not obligatorily) associated with other imitation and975

utilization behaviors [188]. Notably, more frequent976

occurrence of near-CI than of adherent-CI has been977

recently reported also in non-demented Parkinsons978

disease patients [189], whereas the reverse pattern979

has been observed in Parkinsons disease dementia980

[190].981

In his original description, Mayer-Gross [23]982

ascribed CI to a disturbance in performing spatial983

movements of hands and fingers correctly. In later984

years, interpretative accounts of CI clustered within985

two main streams. The first interpretative framework986

might originate from the consideration that CI could987

occur when patients who are unable to structure an988

empty space look for a reference point to solve dif-989

ficult constructional dilemmas [191]. Following this990

suggestion, in recent years CI has been considered as991

a compensatory behavior implemented to overcome992

deficits of visuospatial skills or of visuospatial work-993

ing memory [192]. In this compensation account,994

patients with impaired ability to represent the model,995

or to hold its representation for the need of repro-996

ducing it, might tend to close-in to the model in the997

attempt to reduce their difficulties. Such a compen-998

sation hypothesis received support by the finding999

that complex models could increase severity of CI1000

in AD patients [192], and by the observation that AD1001

patients with CI showed more severe impairments on1002

several visuospatial tests, compared to AD patients1003without CI [193].1004

The alternative account might be traced back to the 1005

hypothesis that CI might represent a primitive reflex 1006

leading patients to be strongly attracted by the model 1007

and be unable to detach from it [24]. In this attrac- 1008

tion framework, CI would represent a default 1009

behavior released by defects of attentional-executive 1010

abilities [194, 195]. In analogy with the compensa- 1011

tion hypothesis, the attraction hypothesis predicts 1012

that CI would be enhanced in dealing with complex 1013

models because they likely imply high attentional 1014

load, and reduce available resources for monitor- 1015

ing graphic productions. However, this hypothesis 1016

specifically envisages strong correlations between CI 1017

and frontal/executive dysfunction, and foresees that 1018

CI can be triggered by high attention-demanding task 1019

conditions. Several converging pieces of evidence 1020

supported this account, since a significant correla- 1021tion between CI and frontal/executive dysfunction 1022

has been reported in MCI [196], AD [191], VaD [25], 1023

and Parkinsons disease [189], and in such diseases 1024

copying geometrical figures in dual-task conditions 1025

enhanced CI [25, 184, 197]. A first attempt at iden- 1026

tifying the neural correlates of CI in AD patients 1027

would point to bilateral orbito-frontal cortex as the 1028

area in which atrophy was significantly associated to 1029

presence of CI [198]. 1030

Most available evidence would thus suggest that, 1031

independently from the diagnosis, patients sharing 1032

similar cognitive impairment might produce CI, and 1033

this attempt at identifying cognitive and neural corre- 1034

lates of specific errors might provide further insights 1035

on drawing disorders in dementia. 1036

CONCLUDING REMARKS 1037

It is interesting to note, at the end of this review, 1038

that several patients reported by Kleist [27] and 1039

Mayer Gross [23] in the earliest studies on construc- 1040

tional apraxia and drawing disorders were affected by 1041

dementia. It is also worth of note that, in one of the 1042

first studies aiming to investigate with standardized 1043

tests the patterns of neuropsychological impairment 1044

shown by various diagnostic groups of dementia, 1045

Gainotti et al. [199] showed that AD patients were 1046

particularly impaired on memory and drawing tasks. 1047

In another study, Villa et al. [200] showed that 1048

the coexistence of mental deterioration was more 1049

strongly associated with drawing disabilities than lat- 1050

erality of lesion, or intrahemispheric locus of lesion 1051

in focal brain-damaged patients. More recent investi- 1052gations have followed two complementary strategies 1053


16/23


to clarify incidence and qualitative aspects of draw-1054

ing disabilities observed in AD and other forms of1055

dementia. The first strategy consisted in adopting1056

short drawing tasks that could be particularly appro-1057

priate to investigate drawing disorders in demented1058

patients. The second strategy consisted in taking into1059

account the nature and the qualitative aspects of1060

drawing disorders that could facilitate the differential1061

diagnosis between AD and the other clinical forms of1062

dementia.1063

The most popular task adopted according to1064

the first strategic approach is CDT, which allows1065

obtaining different scores (e.g., graphic difficul-1066

ties, related to elementary motor disorders; stimulus1067

bound responses, resulting from inhibition difficul-1068

ties; conceptual deficits, related to semantic memory1069

disorders; perseverations resulting from frontal lobe1070dysfunctions and visuospatial disorders), and can be1071

administered both under verbal command and in the1072

copy of a pre-drawn clock (see Fig. 1). According1073

to several authors [21, 88, 125, 126, 201, 202], the1074

CDT is able to distinguish different severity levels1075

and clinical forms of dementia, but other authors1076

reported contrasting findings with respect to both1077

issues [65, 123, 128, 131, 158]. These deceiving1078

results are probably due to two main factors. The1079

first is that the CTD is basically a screening test and as1080

such cannot provide a large and diversified amount of1081

clinical information. The second is that many admin-1082

istration procedures and scoring systems have been1083

proposed [20, 21, 5457], but there is no consensus1084

as to which is the most useful for dementia screen-1085

ing or for providing hints at a differential diagnosis1086

among different forms of dementia. It is also possible1087

that different diagnostic criteria and, above all, rate1088

of diagnostic uncertainty might have biased some of1089

the findings reported in the present review, and this1090

concern might only be overcome by gathering fur-1091

ther well-designed and accurate studies on drawing1092

disorders in the different forms of dementia.1093

More substantial results have been obtained fol-1094

lowing the second research strategy, because several1095

studieshaveshownthatthediscrepancybetweencom-1096

mand and copy condition (in the CDT or in other1097

drawing tasks) is greater in forms of dementia, such1098

as the subcortical vascular dementia [127, 128] or the1099

bvFTD [129, 130, 134136], which are characterized1100

by a prominent deficit of the executive and control1101

functions. Furthermore, the drawing impairments in1102

AD andFTD patients could be based on differentcog-1103

nitive andneuroanatomical mechanisms in agreement1104with the different prevalence of lesions in these two1105

diseases [134, 135]: in AD patients poor figure copy 1106

was associated with performance on spatial percep- 1107

tion and attentional tasks, and correlated significantly 1108

with volumes in the right parietal cortex; in FTD 1109

patients, instead, performance on figure copy cor- 1110

related significantly with scores on spatial planning 1111

and working memory tasks, and was connected with 1112

right dorsolateral prefrontal cortex volumes.Very dif- 1113

ferent from the drawing disorders of patients with 1114

bvFTD are those of patients with semantic dementia, 1115

because both obtain good performances on copying 1116

tasks,butthelatterareselectivelyimpairedindrawing 1117

from memory the typical shape of common objects. 1118

In fact, when SD patients were given the names of 1119

concrete objects and asked to produce the correspond- 1120

ing designs, these drawings were characterized by 1121

a loss of the distinctive features of the represented 1122objects, resulting from disruption of the correspond- 1123

ing conceptual representations, stored in the right and 1124

left anterior temporal lobes [19, 203]. Copying tasks 1125

are, therefore, substantially spared in degenerative 1126

diseases mainly involvingthe frontal (bvFTD)or tem- 1127

poral(SD)corticesandmightbemildlyimpairedinthe 1128

early stages of standard forms of AD, in which atro- 1129

phy prevails in the medial temporal lobes, although 1130

neuroimaging studies repeatedly showed a pattern 1131

of reduced metabolism in the temporo-parietal areas 1132

[204, 205]. A much more severe impairment in copy- 1133

ing tasks is observed in early-onset AD patients and 1134

in PCA variant, in which severe constructive defects 1135

are usually associated with progressive visual-spatial 1136

disorders, including Balint syndrome and unilateral 1137

spatial neglect [100, 101]. A very similar pattern of 1138

drawing impairment is found in LBD, in which visu- 1139

ospatial difficulties are often early and prominent and 1140

contribute to the disproportionate impairment in con- 1141

structionaltasksobservedinthesepatients[151153]. 1142

We can, therefore, conclude that the pattern of 1143

drawing disorders observed in AD and in other forms 1144

of dementia can be explained if we take into account 1145

the main functions of the neural networks involved 1146

in these forms of degenerative diseases. Among 1147

the peculiar drawing patterns that are observed in 1148

the moderate-to-severe forms of dementia the most 1149

important is certainly the closing-in symptom (see 1150

Figs. 2 and 3), which is significantly more frequent 1151

and more severe in AD than in vascular forms of 1152

dementia [24, 25], because AD patients tend to put 1153

the pencil directly over the model, or to overlap the 1154

lines of the model with those of the copy (adherent- 1155

CI), whereas VaD patients more often draw in close 1156proximity to the model without overlapping onto it 1157


17/23


(near-CI). If the CI might improve the differential1158

diagnosis between AD and VaD, it is less useful in1159

the distinction between AD and FTD, because the1160

role of the frontal lobes in its production is still under1161

scrutiny. From this point of view, it is important1162

that recent neuroanatomical data demonstrated the1163

selective association of atrophy in the orbito-frontal1164

areas, implied in inhibiting primitive reflexes, with1165

the closing-in phenomenon [198]. Even if several1166

neuropsychological and pathophysiological aspects1167

of drawing disorders in ADandother formsof demen-1168

tia require further clarification, the clinical interest1169

of drawing tasks in the differential diagnosis among1170

different forms of dementia can, therefore, be con-1171

sidered as firmly established, particularly taking into1172

account analysis of drawing errors more than overall1173

accuracy in drawing.1174

DISCLOSURE STATEMENT1175

Authors disclosures available online (http://j-alz.1176

com/manuscript-disclosures/16-0009r2).1177

REFERENCES1178

[1] Ettlinger G, Warrington E, Zangwill OL (1957) A further1179

study of visual-spatial agnosia. Brain80, 335-361.1180[2] Arena R, Gainotti G (1978) Constructional apraxia and1181

visuoperceptive disabilities in relation to laterality of cere-1182bral lesions.Cortex14, 463-473.1183

[3] Gainotti G, Tiacci C (1970) Patterns of drawing disability1184in right and left hemispheric patients. Neuropsychologia11858, 379-384.1186

[4] Trojano L, Fragassi NA, Chiacchio L, Izzo O, Izzo G,1187Di Cesare G, Cristinzio C, Grossi D (2004) Relationships1188between constructionaland visuospatial abilities in normal1189subjects and in focal brain-damaged patients. J Clin Exp1190

Neuropsychol26, 1103-1112.1191[5] Serra L, Fadda L, Perri R, Spano B, Marra C, Castelli1192

D, Torso M, Makovac E, Cercignani M, Caltagirone C,1193Bozzali M (2014) Constructional apraxia as a distinc-1194tive cognitive and structural brain feature of pre-senile1195Alzheimers disease.J Alzheimers Dis38, 391-402.1196

[6] Warrington EK, James M, Kinsbourne M (1966) Drawing1197disability in relation to laterality of cerebral lesion.Brain119889, 53-82.1199

[7] Hecaen H, Assal G (1970) A comparison of construc-1200tive deficits following right and left hemispheric lesions.1201

Neuropsychologia8, 289-303.1202[8] Gainotti G, Miceli G, Caltagirone C (1977). Construc-1203

tional apraxia in left brain-damaged patients: A planning1204disorder?Cortex13, 109-118.1205

[9] Kirk A, Kertesz A (1991) On drawing impairment in1206Alzheimers disease.Arch Neurol 48, 73-77.1207

[10] Pillon B (1981) Troubles visuo-constructifs et methodes1208

de compensation. Resultats de 85 patients atteints de1209lesions cerebrales.Neuropsychologia19, 375-383.1210

[11] Critchley M (1953) The ParietalLobes, HafnerPress, New 1211York. 1212

[12] Piercy M,Hecaen H, Ajuriaguerra J (1960)Constructional 1213apraxia associated with unilateral cerebral lesions-left and 1214right sided cases compared.Brain83, 225-242. 1215

[13] De Renzi E (1982) Disorders of Space Exploration and 1216

Cognition, Wiley, Chichester. 1217

[14] Makuuchi M, Kaminaga T, Sugishita M (2003) Both pari- 1218etal lobes are involved in drawing: A functional MRI 1219study and implications for constructional apraxia. Brain 1220

Res Cogn Brain Res 16, 338-347. 1221

[15] Roth HL, Bauer RM, Crucian GP, Heilman KM 1222(2014) Frontal-executive constructional apraxia: When 1223delayed recall is better than copying. Neurocase 20, 1224283-295. 1225

[16] Price CC, Cunningham H, Coronado N, Freedland A, 1226Cosentino S, Penney DL, Penisi A, Bowers D, Okun MS, 1227Libon DJ (2011). Clock drawing in the montreal cognitive 1228assessment: Recommendations for dementia assessment. 1229

Dement Geriatr Cogn Disord31, 179-187. 1230

[17] Gainotti G, Silveri MC, Villa G, Caltagirone C (1983) 1231Drawing objects from memory in aphasia. Brain 106, 1232613-622. 1233

[18] Kirk A, Kertesz A (1989) Hemispheric contributions to 1234drawing.Neuropsychologia27, 881-886. 1235

[19] Bozeat S, Lambon Ralph MA, Graham KS, Patterson K, 1236Wilkin H, Rowland J, Hodges JR (2003) A duck with 1237four legs: Investigating the structure of conceptual knowl- 1238edge using picture drawing in semantic dementia. Cogn 1239

Neuropsychol20, 27-47. 1240[20] Agrell B, Dehlin O (2012) The clock-drawing test. 1998. 1241

Age Ageing 41(Suppl 3), 41-45. 1242[21] Freedman MI, Leach L, Kaplan E, Winocur G, Shulman 1243

KJ, Delis DC (1994) Clock Drawing, Oxford University 1244

Press, Oxford. 1245

[22] Trojano L, Conson M (2008). Visuospatial and visuocon- 1246structive deficits. InHandbook of Clinical Neurology, Vol. 124788, Goldenberg G, Miller BL, eds. Elsevier, Amsterdam, 1248pp. 373-392. 1249

[23] Mayer Gross W (1935)Some observation on apraxia. Proc 1250R Soc Med28, 1203-1212. 1251

[24] Gainotti G (1972) A quantitative study of the closing- 1252in symptom in normal children and in brain-damaged 1253patients.Neuropsychologia10, 429-436. 1254

[25] De Lucia N, Grossi D, Trojano L (2014) The genesis of 1255closing-in in Alzheimer disease and vascular dementia: A 1256comparative clinical and experimental study. Neuropsy- 1257chology28, 312-318. 1258

[26] Poppelreuter W (1917) Die Psychischen Schaedigungen 1259durch Kopfschuss im Kriege 1914-1916, Vol. 1, Voss, 1260Leipzig. 1261

[27] Kleist K (1934)Gehirnpathologie, Barth, Leipzig. 1262[28] Gainotti G (1985) Constructional apraxia. InHandbook of 1263

Clinical Neurology, Vol. 45, JAM Fredericks, ed.,Elsevier, 1264Amsterdam, pp. 491-506. 1265

[29] Benton AL (1962) The visual retention test as a construc- 1266tional praxis task.Confin Neurol22, 141-155. 1267

[30] Arrigoni C, De Renzi E (1964) Constructional apraxia and 1268hemispheric locus of lesion.

drawing disorders in alzheimer's disease

Documents